Compact and efficient ultraviolet (UV) laser sources in the wavelength range 200-400 nm are desirable for many applications. Because ultraviolet lasers can be focused to smaller spot sizes than longer-wavelength light sources, they are in demand for high resolution patterning and for drilling tiny holes in miniature circuit boards. In the semiconductor industry, ultraviolet lasers can be used for the detection of foreign particles in clean rooms. This helps to reduce problems in the miniaturization of patterns associated with the integration of semiconductor devices. Similarly, short-wavelength lasers are advantageous for the examination of wafer surfaces in the semiconductor industry. In semiconductor industry manufacturing processes, ultraviolet lasers are widely used in lithography, although existing systems are neither compact nor efficient and are based on excimer laser technology. In consumer electronics, short wavelength lasers can be used for the formation of the recording pits that define high density recording.
Since short-wavelength radiation is easily absorbed by most materials, another application is the detection and classification of materials and substances in security and defense. One technique used in such laser applications is called fluorescence spectroscopy and is based on the ability of ultraviolet light to excite molecules of target materials and produce detectable and distinct wavelength-shifted emission spectra. Another security-related application is the detection of contaminants in water supplies and other materials and also the treatment of water to eliminate biohazards. The properties of high absorption of UV beams and tight focusing make them very useful for nanotechnology and biophotonics. One particular analytical technique enabling these fields is mass spectroscopy. A known use of UV lasers, e.g. at a wavelength of 266 nm is to assist in desorption of analyte molecules from a sample.
Traditionally, ultraviolet lasers have been obtained from a bulky and high-cost gas lasers. An important class of such lasers is called “excimer” (excited dimer) lasers that employ a mixture of a reactive gas (such as F2 or C12) and an inert gas (such as Kr, Ar or Xe) as an active medium. The gas mixtures, when electrically excited, produce a pseudo-molecule, or “dimer”, with an energy level configuration that allows the generation of specific ultraviolet laser wavelengths. As mentioned, the inefficiency, large size, and significant cost of such lasers prevent them from being used in many applications. In summary, there is a substantial and growing demand for compact and efficient ultraviolet laser sources that provide enhanced levels of technical performance, reliability, and cost efficiency. Ideally, one would prefer to obtain a compact ultraviolet laser source directly from a semiconductor device. However, the closest candidates that are based on GaN material system are only able to produce light with wavelengths longer than 370 nm. No other viable semiconductor material has yet been developed to provide lasers with shorter wavelengths.
Another, newer laser platform that has been used for obtaining ultraviolet laser light is based on diode-pumped solid-state (DPSS) lasers which are frequency converted from infrared wavelengths into ultraviolet wavelength via nonlinear optics. This laser platform is most commonly based on Nd-doped solid-state crystals that can produce efficient output at several infrared wavelengths, such as 1064 nm, 946 nm, 914 nm, 1340 nm. The laser wavelength of 1064 nm is a dominant laser wavelength for such gain materials as Nd:YAG, Nd:YVO4, and Nd:GdVO4. The nonlinear optical crystals, such as bulk nonlinear materials as KTP(potassium titanyl phosphate) or LBO (lithium borate) can convert the infrared wavelength of 1064 nm into the green wavelength of 532 nm via a second harmonic generation process (SHG). Further, the nonlinear conversion processes called third-harmonic generation (THG) and fourth-harmonic generation allow obtaining the ultraviolet wavelengths of 355 nm and 266 nm, respectively. Ultraviolet laser products, based on solid-state laser platform are now available from commercial laser manufacturers.
To understand better the design parameters for generating UV light, one can refer to the book by W. P. Risk, T. R. Gosnell and A. V. Nurmikko, “Compact Blue-Green Lasers”, Cambridge University Press (2003), at page 50. The process of generating UV light consists of two nonlinear processes both of which are sum-frequency generation (SFG) processes, described asω1+ω2=ω3 in frequency or   (1)1/λ1+1/λ2=1/λ3 in wavelength.   (2)
The first step in generating UV light is generating visible (e.g., green) light as1064 nm+1064 nm→532 nm   (3)and then mixing the visible and IR beams, as1064 nm+532 nm→355 nm   (4)or doubling the frequency of the visible beam,532 nm+532 nm→266 nm   (5)The power P3 of the generated sum-frequency beams can be estimated using the equation:
                                          P            3                    =                                                    32                ⁢                                  π                  2                                ⁢                                  d                  eff                  2                                                                              ɛ                  0                                ⁢                                  cn                  3                  2                                ⁢                                  λ                  1                                ⁢                                  λ                  2                                ⁢                                  λ                  3                                                      ⁢                          P              1                        ⁢                          P              2                        ⁢            lh                          ,                            (        6        )            where P1, P2, and P3 denote the power of the optical beams participating in the nonlinear process, λ1, λ2, and λ3 denote wavelengths of these beams, deff is the effective nonlinear coefficient for the nonlinear crystal, l is the length of the nonlinear crystal, h is the Boyd-Kleinmann's function that depends on the degree of focusing of the fundamental beam, c is the speed of light, εo is the dielectric constant of a vacuum, and n3 is the dielectric constant for the sum-frequency beam.
Equation (6) is useful in understanding the limitations of existing platforms for UV laser sources. Despite obvious advantages in efficiency and size of DPSS UV laser sources compared to gas lasers, their size, efficiency, and cost are still not satisfactory for many newer applications, especially when a UV laser source is designed to be part of a compact, portable instrument. One of the primary design limitations is the inefficiency of multiple nonlinear processes, which are required to convert the fundamental infrared laser beam into the ultraviolet beam.
To illustrate the prior art way of obtaining UV laser radiation, one can refer to U.S. Pat. No. 7,016,389. This patent describes architectures for improving nonlinear frequency tripling from the fundamental wavelength of 1064 nm into the third harmonic wavelength of 355 nm. To improve the conversion efficiency into the UV, both the second-harmonic generation (SHG) process, generating green, and the third-harmonic generation (THG) process, generating UV, are done inside the laser cavity to take advantage of the high circulating power at the fundamental wavelength. Further improvements in nonlinear conversion efficiency are achieved via Q-switching that raises the intracavity peak power at the fundamental laser wavelength.
The nonlinear materials listed in the U.S. Pat. No. 7,016,389 as being suitable for frequency conversion processes are well known bulk nonlinear crystals LBO, KTP, KNbO3, CLBO, BBO which achieve nonlinear conversion via a process called birefringent phase-matching. Of these materials, LBO can be used for generating both green and UV wavelength light. The limitation of LBO is its low nonlinearity with an effective nonlinear coefficient of under 1 pm/V for both SHG and THG processes. This limitation makes it essentially mandatory to employ complex and costly laser architectures to increase peak power and nonlinear conversion efficiencies. Using any of the nonlinear crystals listed above does not lead to an efficient and low-cost UV platform since the most efficient crystal in that list, KNbO3, can be useful for generating blue wavelengths but cannot be used for efficient generation of the green and UV wavelengths due to the high walkoff and poor reliability problems. The KTP crystal, which has deff˜3.5 pm/V for SHG conversion into the green wavelength, has reliability limitations (known as gray tracking) and is used primarily in low power green lasers. The borates (BBO, CLBO, LBO) are useful for conversion into the UV wavelength but are limited in their efficiency. Thus, it is difficult to find a combination of two nonlinear crystals that ensure efficient UV generation in a compact, low-cost architecture.
Similar limitations apply to approaches described in other patents. U.S. Pat. No. 6,002,695 and U.S. Pat. No. 6,697,391, describe high-efficiency frequency tripling or quadrupling with both nonlinear processes taking place inside the laser cavity. LBO crystals are mentioned as the preferred material for both nonlinear conversions. In addition, U.S. Pat. No. 6,697,391 suggests the use of other bulk nonlinear converters, including KTP, BBO, ADP, CBO, DADA, DADP, DKDP, and others. The platforms described in U.S. Pat. Nos. 6,002,695 and 6,697,391 are somewhat efficient but cannot be built in a compact and low-cost package due to multiple design constraints such as keeping the fixed and optimized beam sizes in both the solid-state gain crystal and in the nonlinear crystals.
A number of prior art workers chose external nonlinear conversion to improve reliability and simplicity of the UV laser system. An example of such work is U.S. Pat. No. 6,157,663. The preferred nonlinear converter for both green (SHG) and UV (THG) is LBO. While this system has advantages due to decoupling of nonlinear conversion processes from the laser cavity, it is still large and costly due to the inefficiency of nonlinear conversions in LBO. The cost is manifested in the number of laser components such as focusing lenses, temperature controls, etc. and in the size of the overall system.
Recently, there have been efforts to find a compact and low cost platform for UV laser sources. One approach is based on the so-called microchip solid-state laser architecture with passive Q-switching. A laser system based on this approach is described in the U.S. Pat. No. 6,373,864. This patent had been preceded by a number of research papers on microchip laser platform with or without Q-switching (see the list of references in U.S. Pat. No. 6,373,864). The enabling feature of this laser platform is a compact, monolithic resonator defined by the solid-state gain crystal (e.g., Nd:YAG) and saturable absorber medium (e.g., Cr4+:YAG). The system is compact, low-cost, and delivers high peak power in sub-nanosecond pulses, which make it easier to achieve high nonlinear conversion efficiency. However, this approach is still limited in power, efficiency, and reliability due to the choice of traditional nonlinear converters: KTP, BBO, or LBO.
Even more recently, attempts to improve nonlinear conversions with newer nonlinear crystals have been made. U.S. Pat. No. 6,741,620, suggests the use of a PPKTP (periodically poled KTP) nonlinear crystal for efficient frequency doubling into the visible and then using CLBO for converting into UV. The advantage of PPKTP is its high nonlinear coefficient deff˜9 pm/V (compared to ˜3.5 pm/V for KTP and <1 pm/V for LBO). This advantage is important because SHG power scales with the square of deff (see Eq. (6)). However, the design of U.S. Pat. No. 6,741,620 is neither compact nor low-cost and, in addition, PPKTP is based on KTP material and it also suffers from the same reliability (gray tracking) limitations as bulk KTP. Therefore, it has very limited potential for a scalable, low-cost, and reliable UV platform.
K. Mizuuchi, A. Morikawa, T. Sugita, K. Yamamoto, N. Pavel, and T. Taira (“Continuous-wave ultraviolet generation at 354 nm in a periodically poled MgO:LiNbO3 by frequency tripling of a diode end-pumped Nd:GdVO4 microlaser,” Applied Physics Letters, vol. 85, p. 3959 (2004)) demonstrated that periodically poled MgO-doped LiNbO3 (PPMgOLN) can be used for both efficient SHG (into the green) and THG (into the UV). This material has one of the highest nonlinear coefficients (deff˜16 pm/V) of all nonlinear materials. This research paper described the concept of using PPMgOLN for UV generation. However, this paper did not provide a design for a low-cost, robust UV laser system, suitable for commercial production.
The potential of periodically poled materials based on LiNbO3 and LiTaO3 has been recognized for some time. However, the efforts focused on periodic poling of traditionally grown (congruent) versions of these materials. While the fabrication of these periodically poled materials has been demonstrated, they suffered from photo-refractive degradation during laser operation unless the temperature of these materials is raised to >150° C. This is a serious limitation in adopting PPLN (periodically poled lithium niobate) and PPLT (periodically poled lithium tantalate) in commercial laser systems. Therefore, doping the material with such impurities as MgO and ZnO during the crystal growth to suppress photo-refractive mechanisms (T. Volk, N. Rubinina, M. Wöhlecke, “Optical-damage-resistant impurities in lithium niobate,” Journal of the Optical Society of America B, vol. 11, p. 1681 (1994)) was proposed. Finally, growing the crystals with a high degree of stoichiometry with or without dopants has been proposed as another method to suppress photo-refractive damage (Y. Furukawa, K. Kitamura, S. Takekawa, K. Niwa, H. Hatano, “Stoichiometric Mg:LiNbO3 as an effective material for nonlinear optics,” Optics Letters, vol. 23, p. 1892 (1998)).
However, it has been recognized that MgO and ZnO-doped and stoichiometric LiNbO3 and LiTaO3 are very different materials from their congruent counterparts and their modified ferroelectric properties make these materials exceedingly difficult to pole into the short-periods, several-micron-length domains, required for frequency conversion into the visible and UV spectral ranges. Therefore, periodic poling of such materials has not been done in a production environment and suggestions concerning their technological potential (such as in the paper by Mizuuchi et al.) are quite rare.